A hierarchical multi-UAV task allocation method based on dominance rough sets

Abstract

The invention belongs to the field of unmanned aerial vehicle task allocation, and specifically relates to a hierarchical multi-unmanned aerial vehicle task allocation method based on dominance rough sets; the method comprises: normalizing the attribute values ​​of unmanned aerial vehicles; constructing a total utility objective function according to the normalized attribute values ​​of unmanned aerial vehicles; fusing the information of unmanned aerial vehicles according to the total utility objective function to obtain a total attribute fusion value; defining the dominance relationship between unmanned aerial vehicles according to the total attribute fusion values ​​of different unmanned aerial vehicles, and obtaining the dominance class of unmanned aerial vehicles according to the dominance relationship between unmanned aerial vehicles; calculating the dominance degree between unmanned aerial vehicles according to the dominance class of unmanned aerial vehicles; calculating the overall dominance degree of unmanned aerial vehicles according to the dominance degree between unmanned aerial vehicles; allocating task targets according to the overall dominance degree of unmanned aerial vehicles corresponding to the task targets, and obtaining the allocation result of the task targets; the invention can improve the efficiency of multi-unmanned aerial vehicle task allocation without relying on too many parameters, and has high practicality.

Description

Technical Field

[0001] The present invention belongs to the field of unmanned aerial vehicle task allocation, and in particular relates to a hierarchical multi-unmanned aerial vehicle task allocation method based on dominant rough sets. Background Art

[0002] With the rapid development of information technology, especially artificial intelligence technology, drones are playing an important role in many fields. Drones have the advantages of small size, low cost, high stealth, and good maneuverability. They are suitable for performing complex tasks in harsh or dangerous environments. As the complexity of the task increases, a single drone is often difficult to achieve and requires multiple drones to work together. However, with the increasing complexity of drones and mission objectives, drone task allocation has become more and more complex, making it increasingly difficult to solve the multi-drone task allocation problem. Proposing efficient multi-drone task allocation algorithms and models has become a research hotspot for scholars.

[0003] Currently, the research on multi-UAV task allocation mainly focuses on optimization-based methods, including auction-based algorithms, contract net-based methods, and meta-heuristic algorithms such as swarm intelligence algorithms and evolutionary algorithms.

[0004] Traditional auction-based algorithms and contract network-based algorithms can effectively solve the problem of multi-UAV task allocation, but as the complexity of the task increases, the calculation time usually grows exponentially. Algorithms based on swarm intelligence and evolutionary algorithms can achieve good performance when solving small-dimensional or low-dimensional problems, but when solving large-scale or high-dimensional problems, swarm intelligence-based algorithms and evolutionary algorithms are prone to fall into premature or local optimality, which makes the algorithm performance worse; and swarm intelligence-based algorithms and evolutionary algorithms are dependent on parameter settings. At present, no researchers have used rough set theory and methods to solve the problem of multi-UAV task allocation. The present invention uses the advantage rough set theory and combines human cognition to design a multi-granularity progressive decomposition mechanism "from coarse-grained to fine-grained", establish a hierarchical multi-UAV task allocation algorithm, solve the problems of high computational complexity and dependence on parameter settings in the prior art, and improve the efficiency of multi-UAV task allocation. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a hierarchical multi-UAV task allocation method based on dominant rough sets, which includes:

[0006] S1: Obtain a set of task targets to be assigned and a set of drones to be assigned;

[0007] S2: normalizing the drone attribute values ​​of all drones in the set of drones to be assigned corresponding to all task targets in the set of task targets to be assigned;

[0008] S3: Set the weights of the drone attributes; construct the total utility objective function based on the weights of the drone attributes and the normalized drone attribute values;

[0009] S4: Perform information fusion on the UAV according to the total utility objective function to obtain the total attribute fusion value;

[0010] S5: define the advantage relationship between drones according to the total attribute fusion values ​​of different drones, and obtain the advantage class of drones according to the advantage relationship between drones;

[0011] S6: Calculate the dominance between drones according to the dominance categories of drones;

[0012] S7: Calculate the overall dominance of the drones according to the dominance between the drones;

[0013] S8: Allocate the mission targets according to the overall dominance of the UAVs corresponding to the mission targets to obtain the allocation result of the mission targets; the UAVs process the mission targets according to the allocation result;

[0014] S9: Update the set of task targets to be assigned and the set of drones to be assigned;

[0015] S10: Determine whether the number of task targets in the set of task targets to be assigned is zero. If it is zero, the task assignment is completed; otherwise, return to step S4.

[0016] Preferably, the formula for normalizing the drone attribute values ​​is:

[0017]

[0018] in, represents the normalized i-th drone x i In attribute a k The drone attribute value on represents the i-th drone x before normalization i In attribute a k The drone attribute value on Indicates that in attribute a k The minimum drone attribute value on Indicates that in attribute a k The maximum drone attribute value on .

[0019] Preferably, the total utility objective function is:

[0020]

[0021] stx im ∈{0,1}

[0022]

[0023]

[0024]

[0025] Among them, λ k ,k∈{1,2,3,4,5} represents the weight of the kth attribute, a k , k∈{1,2,3,4,5} represents the kth attribute, M represents the number of mission targets, N represents the number of drones, x im Indicates whether the i-th UAV completes the m-th mission objective.

[0026] Preferably, the formula for information fusion of drones is:

[0027]

[0028] Among them, f A (x i ) represents the total attribute fusion value, λ k represents the kth attribute a k The weight of represents the i-th drone x i In attribute a k The drone attribute value on a k represents the kth attribute, and A represents the attribute set.

[0029] Preferably, the advantage relations between drones and the advantage classes of drones are:

[0030]

[0031]

[0032] in, represents the i-th drone x i and the jth drone x j The advantage relationship between A (x i ) represents the i-th drone x i The total attribute fusion value, f A (x j ) represents the jth drone x j The total attribute fusion value of , U represents the drone set, A represents the attribute set, Indicates that it is greater than or equal to the i-th drone x i The drone advantage class with the total attribute fusion value.

[0033] Preferably, the formula for calculating the dominance between drones is:

[0034]

[0035] Among them, D A (x i ,x j ) represents the i-th drone x i With the jth drone x j The dominance of Indicates that it is greater than or equal to the i-th drone x i The total attribute fusion value of the drone advantage class, Indicates that it is greater than or equal to the jth drone x j The total attribute fusion value of the UAV advantage class, U represents the UAV set.

[0036] Preferably, the formula for calculating the overall dominance of the drone is:

[0037]

[0038] Among them, D A (x i ) represents the i-th drone x i The overall dominance, D A (x i ,x j ) represents the i-th drone x i With the jth drone x j The dominance of , U represents the set of drones.

[0039] Preferably, the process of allocating task objectives includes:

[0040] Select the UAV with the highest overall superiority for a mission target. If its overall superiority for other mission targets is not the highest, assign the mission target to the UAV with the highest overall superiority.

[0041] Select the UAV with the highest overall advantage for a mission target. If its overall advantage for other mission targets is also the highest, then assign the mission target with the highest overall advantage to the UAV.

[0042] The beneficial effects of the present invention are as follows: the present invention uses the dominance rough set theory to fuse the acquired information, define the corresponding dominance relationship, calculate the overall dominance of all the drones to be assigned relative to each task target, and assign the task targets according to the overall dominance; the present invention combines human cognition to design a multi-granularity progressive decomposition mechanism "from coarse granularity to fine granularity" and establish a hierarchical multi-drone task allocation algorithm. Compared with the existing technology, the present invention does not rely on additional parameters, and the obtained allocation results have a larger total utility and a shorter running time, and are highly efficient. When the number of drones and the number of tasks increase, the advantages of total utility and running time also expand, and the practicality is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a flow chart of the hierarchical multi-UAV task allocation method based on dominant rough sets in the present invention.

[0044] Figure 2 This is a framework diagram of the hierarchical multi-UAV task allocation method based on advantage rough sets in the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] This paper proposes a hierarchical multi-UAV task allocation method based on advantage rough sets. Figure 1 As shown, the method includes the following contents:

[0047] S1: Obtain a set of task targets to be assigned and a set of drones to be assigned.

[0048] Obtain drone information and mission target information, obtain the mission targets and drones to be assigned, the number of drones to be assigned is greater than or equal to the number of mission targets, and it is stipulated that in one assignment process, one drone only handles one mission target.

[0049] S2: Normalize the drone attribute values ​​of all drones in the set of drones to be assigned corresponding to all task targets in the set of task targets to be assigned.

[0050] The drone attribute values ​​corresponding to each task target to be assigned are normalized. The drone attributes include the distance between the drone and the task target, that is, the Euclidean distance between the coordinates of the drone and the task target, the time for the drone to complete the task target, the cost-effectiveness of the drone to complete the task target, the drone's ability to complete the task target, and the loss of the task target to the drone. The formula for normalizing the drone attribute values ​​is:

[0051]

[0052] in, represents the normalized i-th drone x i In attribute a k The drone attribute value on represents the i-th drone x before normalization i In attribute a k The drone attribute value on Indicates that in attribute a k The minimum drone attribute value on Indicates that in attribute a k The maximum drone attribute value on .

[0053] S3: Set the weights of the drone attributes; construct the total utility objective function based on the weights of the drone attributes and the normalized drone attributes.

[0054] Set the weight λ of the drone attribute k ,k∈{1,2,3,4,5}; The weight of the drone attribute can be set manually according to the importance of the attribute. Preferably, for the convenience of calculation, the present invention sets the attribute weight value to λ 1 =λ 2 =1,λ 3 =λ 4 =λ 5 =-1.

[0055] The total utility objective function is:

[0056]

[0057] stx im ∈{0,1}

[0058]

[0059]

[0060]

[0061] Among them, λ k ,k∈{1,2,3,4,5} represents the weight of the kth attribute, ak , k∈{1,2,3,4,5} represents the kth attribute, M represents the number of mission targets, and N represents the number of drones; x im Indicates whether the i-th UAV completes the m-th mission target. If the i-th UAV completes the m-th mission target, then x im =1, otherwise, x im =0.

[0062] Solve the total utility objective function. When all tasks are assigned, the final result of the task assignment can be obtained. The process of solving the total utility objective function is:

[0063] S4: Perform information fusion on the UAVs according to the total utility objective function to obtain the total attribute fusion value.

[0064] For an ordered information system S = (U, A, V, f), where U is the set of drones, A is the set of attributes, V is the value range of all attributes, and f is the mapping of drones to attribute values, based on the ordered information system, the information fusion of drones is performed according to the weights of drone attributes. The formula for information fusion of drones is:

[0065]

[0066] Among them, f A (x i ) represents the total attribute fusion value, λ k represents the kth attribute a k The weight of represents the i-th drone x i In attribute a k The drone attributes on k represents the kth attribute, and A represents the attribute set.

[0067] S5: Define the dominance relationship between drones according to the total attribute fusion values ​​of different drones, and obtain the dominance class of drones according to the dominance relationship between drones.

[0068] The dominance relationships between drones and the dominance classes of drones are:

[0069]

[0070]

[0071] in, represents the i-th drone x i and the jth drone x j The advantage relationship between them is that the i-th drone x i The total attribute fusion value of is less than or equal to the jth drone x jThe total attribute fusion value of A (x i ) represents the i-th drone x i The total attribute fusion value, f A (x j ) represents the jth drone x j The total attribute fusion value of , U represents the drone set, A represents the attribute set, Indicates that it is greater than or equal to the i-th drone x i The UAV advantage class with a total attribute fusion value greater than or equal to the i-th UAV x i A collection of drones with a total attribute fusion value.

[0072] S6: Calculate the dominance between drones according to their dominance categories.

[0073] The formula for calculating the dominance between drones is:

[0074]

[0075] Among them, D A (x i ,x j ) represents the i-th drone x i With the jth drone x j The dominance of Indicates that it is greater than or equal to the jth drone x j The total attribute fusion value of the UAV advantage class, U represents the UAV set.

[0076] Then, the dominance matrix between drones is:

[0077]

[0078] S7: Calculate the overall dominance of the drones based on the dominance between the drones.

[0079] The formula for calculating the overall dominance of the drone is:

[0080]

[0081] Among them, D A (x i ) represents the i-th drone x i The overall advantage of .

[0082] S8: Allocate the mission objectives according to the overall dominance of the UAVs corresponding to the mission objectives to obtain the allocation result of the mission objectives; the UAV processes the mission objectives according to the allocation result.

[0083] The process of assigning task objectives includes:

[0084] All drones under each mission target are ranked from high to low according to the overall superiority of the drone corresponding to the mission target, that is, the higher the overall superiority, the higher the ranking;

[0085] According to the ranking results under different mission objectives, the UAV with the highest overall advantage for a mission objective is selected. If its overall advantage for other mission objectives is not the highest, the mission objective is assigned to the UAV with the highest overall advantage.

[0086] Select the UAV with the highest overall advantage for a mission target. If its overall advantage for other mission targets is also the highest, then assign the mission target with the highest overall advantage to the UAV.

[0087] S9: Update the set of mission targets to be assigned and the set of drones to be assigned.

[0088] Updating the set of task targets to be assigned refers to removing the assigned task targets from the set of task targets, and updating the set of drones to be assigned refers to removing the assigned drones from the set of drones to be assigned.

[0089] S10: Determine whether the number of task targets in the set of task targets to be assigned is zero. If it is zero, the task assignment is completed; otherwise, return to step S4.

[0090] Determine the number of task targets in the set of task targets to be assigned. If the number of task targets is greater than 0, return to step S4 and continue to assign task targets. If the number of task targets is equal to 0, the assignment is completed and the final assignment result of all task targets is obtained. The UAVs assigned with task targets process the task targets according to the assignment results.

[0091] The present invention allocates tasks hierarchically, such as Figure 2 As shown, the first layer is all the drones and mission targets to be assigned; the second layer is the assigned drones and mission targets, and the drones and mission targets to be assigned; the third layer is the redistribution of the unassigned drones and mission targets in the second layer, which is divided into two groups, namely the assigned drones and mission targets, and the drones and mission targets to be assigned; and so on, until all mission targets are assigned.

[0092] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation modes of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A hierarchical multi-UAV task allocation method based on dominant rough sets, characterized in that: include: S1: Obtain a set of task targets to be assigned and a set of drones to be assigned; S2: normalizing the drone attribute values ​​of all drones in the set of drones to be assigned corresponding to all task targets in the set of task targets to be assigned; S3: Set the weights of the drone attributes; construct the total utility objective function based on the weights of the drone attributes and the normalized drone attributes; the total utility objective function is: s.t.x im ∈{0,1} Among them, λ k ,k∈{1,2,3,4,5} represents the weight of the kth attribute, a k , k∈{1,2,3,4,5} represents the kth attribute, M represents the number of mission targets, N represents the number of drones, x im Indicates whether the i-th UAV has completed the m-th mission objective; S4: According to the total utility objective function, the information of the UAV is fused to obtain the total attribute fusion value; the formula for information fusion of the UAV is: Among them, f A (x i ) represents the i-th drone x i The total attribute fusion value of represents the i-th drone x i In attribute a k The drone attribute value on it, A represents the attribute set; S5: Define the advantage relationship between drones according to the total attribute fusion value of different drones, and obtain the advantage class of drones according to the advantage relationship between drones; the advantage relationship between drones and the advantage class of drones are: in, represents the i-th drone x i and the jth drone x j The advantage relationship between A (x j ) represents the jth drone x j The total attribute fusion value of , U represents the drone set, Indicates that it is greater than or equal to the i-th drone x i The drone advantage class with the total attribute fusion value; S6: Calculate the dominance between drones according to the dominance class of the drones; the formula for calculating the dominance between drones is: Among them, D A (x i ,x j ) represents the i-th drone x i With the jth drone x j The dominance of Indicates that it is greater than or equal to the jth drone x j The drone advantage class with the total attribute fusion value; S7: Calculate the overall dominance of the drones based on the dominance between the drones; the formula for calculating the overall dominance of the drones is: Among them, D A (x i ) represents the i-th drone x i The overall dominance of S8: Allocate the mission targets according to the overall dominance of the UAVs corresponding to the mission targets to obtain the allocation result of the mission targets; the UAVs process the mission targets according to the allocation result; S9: Update the set of task targets to be assigned and the set of drones to be assigned; S10: Determine whether the number of task targets in the set of task targets to be assigned is zero. If it is zero, the task assignment is completed; otherwise, return to step S4.

2. The hierarchical multi-UAV task allocation method based on dominant rough sets according to claim 1 is characterized in that: The formula for normalizing the drone attribute values ​​is: in, represents the normalized i-th drone x i In attribute a k The drone attribute value on represents the i-th drone x before normalization i In attribute a k The drone attribute value on Indicates that in attribute a k The minimum drone attribute value on ak ) indicates that in attribute a k The maximum drone attribute value on .

3. The hierarchical multi-UAV task allocation method based on dominant rough sets according to claim 1 is characterized in that: The process of assigning task objectives includes: Select the UAV with the highest overall superiority for a mission target. If its overall superiority for other mission targets is not the highest, assign the mission target to the UAV with the highest overall superiority. Select the UAV with the highest overall advantage for a mission target. If its overall advantage for other mission targets is also the highest, then assign the mission target with the highest overall advantage to the UAV.

Images